{"id":502,"title":"tensorflow tf检查是否支持gpu","content":"```\n1\\下载Visual Studio最新版本\nhttps:\/\/visualstudio.microsoft.com\/zh-hans\/downloads\/\nhttps:\/\/learn.microsoft.com\/zh-cn\/visualstudio\/releases\/2019\/release-notes\n1\\安装显卡驱动\nhttps:\/\/www.nvidia.com\/download\/index.aspx\n2\\GPU支持硬件型号列表\nhttps:\/\/developer.nvidia.com\/cuda-gpus\n3\\下载 CUDA 工具包\nhttps:\/\/developer.nvidia.com\/cuda-downloads\nCUDA版本选择,依据此链接\nhttps:\/\/docs.nvidia.com\/cuda\/cuda-toolkit-release-notes\/index.html\n\nhttps:\/\/tensorflow.google.cn\/install\/gpu\nGPU驱动下载\nhttps:\/\/www.nvidia.com\/download\/index.aspx?lang=en-us\n\n\nCUDA工具包\nhttps:\/\/developer.nvidia.com\/cuda-toolkit-archive?utm_source=tensorflow.google.cn&utm_medium=referral\ncuDNN各版本\nhttps:\/\/developer.nvidia.com\/rdp\/cudnn-archive?utm_source=tensorflow.google.cn&utm_medium=referral\ncudnn储存在目录下system32\n\n\n\nTensorFlow Version: 2.6.2\n\n版本对应\ntf          cuda            cudnn\n2.0         10.0            7.6\n2.1         10.1            7.6\n2.2         10.1            7.6\n2.3         10.1            7.6\n2.4         11.0            8.0\n2.5         11.2            8.1\n2.6         11.2            8.1\n表格数据参考链接底部\nhttps:\/\/tensorflow.google.cn\/install\/source_windows\n\n\nhttps:\/\/developer.download.nvidia.cn\/compute\/cuda\/12.0.0\/local_installers\/cuda_12.0.0_527.41_windows.exe\n\nhttps:\/\/developer.download.nvidia.cn\/compute\/cuda\/11.0.3\/local_installers\/cuda_11.0.3_451.82_win10.exe\n\nhttps:\/\/developer.download.nvidia.cn\/compute\/cuda\/11.2.1\/local_installers\/cuda_11.2.1_461.09_win10.exe\n```\n\n\n```python\n支持GPU的显卡型号\nhttps:\/\/developer.nvidia.com\/cuda-gpus\n\nfrom distutils.version import LooseVersion\nimport warnings\nimport tensorflow as tf\n\n# Check TensorFlow Version\nassert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer'\nprint('TensorFlow Version: {}'.format(tf.__version__))\n\n# Check for a GPU\nif not tf.test.gpu_device_name():\n    warnings.warn('No GPU found. Please use a GPU to train your neural network.')\nelse:\n    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))\n    \n```","sort":"4","view":"3647","addtime":"1671067508","password":"","tagname":["python"],"sorts":"Python"}